Methods and Systems for Predicting Successful Data Transmission During Mass Communications Across Computer Networks Featuring Disparate Entities and Imbalanced Data Sets Using Machine Learning Models

1. A system for predicting success probabilities for settlement transactions across computer networks featuring disparate entities using machine learning models, the system comprising: cloud-based storage circuitry configured to store: a database comprising a plurality of records, wherein each record of the plurality of records corresponds to a respective, previously transmitted settlement transaction, wherein each record is organized according to a respective transmission time of a plurality of transmission times, and wherein each of the plurality of records indicates a settlement transaction characteristic, a recipient identifier, and a settlement transaction identifier; a database of historical network metrics for each of the plurality of transmission times; a first subset of feature inputs, wherein each feature input of the first subset of feature inputs is based on a respective record of the plurality of records and a historical network metric of the historical network metrics that corresponds to a transmission time of each respective record, and wherein the transmission time of each respective record of each feature input of the first subset of feature inputs corresponds to a first time period; a machine learning model: wherein the machine learning model is trained to classify the first subset of feature inputs as corresponding to one of a plurality of predicted settlement transaction success rates at given times, wherein each feature input of the first subset of feature inputs is based on a respective record of the plurality of records, and a historical network metric of the historical network metrics that corresponds to a transmission time of the respective record; and wherein the machine learning model is validated using a second subset of feature inputs, wherein each feature input of the second subset of feature inputs corresponds to a second time period, and wherein the second time period is after the first time period; cloud-based control circuitry configured to: receive a first feature input and first times for prediction, wherein the first feature input is based on a first pending settlement transaction; input the first feature input into the machine learning model, wherein the first feature input corresponds to a third time period, and wherein the third time period is after the second time period; receive an output from the machine learning model; and cloud-based input/output circuitry configured to: generate for display, on a user interface, a predicted settlement transaction success rate for the first pending settlement transaction at each of the first times based on the output.

2. A method for predicting success probabilities for settlement transactions across computer networks featuring disparate entities using machine learning models, the method comprising: receiving, using control circuitry, a first feature input and first times for prediction, wherein the first feature input is based on a record for a first pending settlement transaction, and wherein the record indicates a settlement transaction characteristic, a recipient identifier, and a settlement transaction identifier for the first pending settlement transaction; inputting, using the control circuitry, the first feature input into a machine learning model, wherein the machine learning model is trained to classify a first subset of feature inputs as corresponding to one of a plurality of predicted settlement transaction success rates at given times, wherein each of the first subset of feature inputs corresponds to a respective, previously transmitted settlement transaction, and wherein each feature input of the first subset of feature inputs is based on a respective record of a plurality of records and a historical network metric that corresponds to a transmission time of the respective record; receiving, using the control circuitry, an output from the machine learning model; and generating for display, on a user interface, a predicted settlement transaction success rate for the first pending settlement transaction at each of the first times based on the output.

3. The method of claim 2 , and wherein the transmission time of each respective record of each feature input of the first subset of feature inputs corresponds to a first time period.

4. The method of claim 3 , further comprising: receiving a second feature input, wherein the second feature input is based on the record for the first pending settlement transaction; inputting the second feature input into an additional machine learning model, wherein the additional machine learning model is trained to predict a settlement transaction success rate at any time, wherein the additional machine learning model comprises: a graph neural network, wherein the graph neural network is trained using historical directed graphs corresponding to the first subset of feature inputs, and a second subset of feature inputs, wherein each feature input of the second subset of feature inputs corresponds to a second time period, and wherein the second time period is after the first time period; and a gradient boosting model, wherein the gradient boosting model is trained on outputs of the graph neural network; receiving an additional output from the additional machine learning model; and generating for display, on the user interface, the predicted settlement transaction success rate for the first pending settlement transaction at any time based on the additional output.

5. The method of claim 4 , further comprising: receiving, at the user interface, a first user request for the predicted settlement transaction success rate for the first pending settlement transaction at each of the first times; in response to receiving the first user request, generating the first feature input; receiving, at the user interface, a second user request for the predicted settlement transaction success rate for the first pending settlement transaction at any time; and in response to receiving the second user request, generating the second feature input.

6. The method of claim 3 , wherein the machine learning model is validated using a second subset of feature inputs, wherein each feature input of the second subset of feature inputs corresponds to a second time period, and wherein the second time period is after the first time period.

7. The method of claim 4 , wherein the first feature input corresponds to a third time period, and wherein the third time period is after the second time period.

8. The method of claim 2 , wherein the first feature input includes classes of features based on a rate of successful settlement transactions, a number of settlement transactions corresponding to the settlement transaction identifier, current operational issued related to the recipient identifier, a number of entities with pending settlement transactions, and probabilistic estimations of successful settlement transactions related to the recipient identifier.

9. The method of claim 2 , wherein the first feature input includes an engineered feature class that comprises a baseline and moving average rate of pending settlement transactions.

10. The method of claim 2 , wherein the machine learning model utilizes a gradient boosting framework that uses tree based learning algorithms.

11. The method of claim 2 , further comprising: determining a feature importance of each feature in the first subset of feature inputs when classifying the first subset of feature inputs as corresponding to one of the plurality of predicted settlement transaction success rates by determining a SHAP (“Shapley Additive explanation”) value for each feature.

12. A non-transitory, computer-readable medium for predicting success probabilities for settlement transactions across computer networks featuring disparate entities using machine learning models, comprising instructions that, when executed by one or more processors, cause operations comprising: receiving a first feature input and first times for prediction, wherein the first feature input is based on a record for a first pending settlement transaction, and wherein the record indicates a settlement transaction characteristic, a recipient identifier, and a settlement transaction identifier for the first pending settlement transaction; inputting the first feature input into a machine learning model, wherein the machine learning model is trained to classify a first subset of feature inputs as corresponding to one of a plurality of predicted settlement transaction success rates at given times, wherein each of the first subset of feature inputs corresponds to a respective, previously transmitted settlement transaction, and wherein each feature input of the first subset of feature inputs is based on a respective record of a plurality of records and a historical network metric that corresponds to a transmission time of the respective record; receiving an output from the machine learning model; and generating for display, on a user interface, a predicted settlement transaction success rate for the first pending settlement transaction at each of the first times based on the output.

13. The non-transitory, computer-readable medium of claim 12 , wherein the transmission time of each respective record of each feature input of the first subset of feature inputs corresponds to a first time period.

14. The non-transitory, computer-readable medium of claim 13 , wherein the instructions further cause operations comprising: receiving a second feature input, wherein the second feature input is based on the record for the first pending settlement transaction; inputting the second feature input into an additional machine learning model, wherein the additional machine learning model is trained to predict a settlement transaction success rate at any time, wherein the additional machine learning model comprises: a graph neural network, wherein the graph neural network is trained using historical directed graphs corresponding to the first subset of feature inputs and a second subset of feature inputs, wherein each feature input of the second subset of feature inputs corresponds to a second time period, and wherein the second time period is after the first time period; and a gradient boosting model, wherein the gradient boosting model is trained on outputs of the graph neural network; receiving an additional output from the additional machine learning model; and generating for display, on the user interface, the predicted settlement transaction success rate for the first pending settlement transaction at any time based on the additional output.

15. The non-transitory, computer-readable medium of claim 14 , wherein the instructions further cause operations comprising: receiving, at the user interface, a first user request for the predicted settlement transaction success rate for the first pending settlement transaction at each of the first times; in response to receiving the first user request, generating the first feature input; receiving, at the user interface, a second user request for the predicted settlement transaction success rate for the first pending settlement transaction at any time; and in response to receiving the second user request, generating the second feature input.

16. The non-transitory, computer-readable medium of claim 13 , wherein the machine learning model is validated using a second subset of feature inputs, wherein each feature input of the second subset of feature inputs corresponds to a second time period, and wherein the second time period is after the first time period.

17. The non-transitory, computer-readable medium of claim 16 , wherein the first feature input corresponds to a third time period, and wherein the third time period is after the second time period.

18. The non-transitory, computer-readable medium of claim 12 , wherein the first feature input includes classes of features based on a rate of successful settlement transactions, a number of settlement transactions corresponding to the settlement transaction identifier, current operational issued related to the recipient identifier, a number of entities with pending settlement transactions, and probabilistic estimations of successful settlement transactions related to the recipient identifier.

19. The non-transitory, computer-readable medium of claim 12 , wherein the machine learning model utilizes a gradient boosting framework that uses tree based learning algorithms.

20. The non-transitory, computer-readable medium of claim 12 , wherein the instructions further cause operations comprising: determining a feature importance of each feature in the first subset of feature inputs when classifying the first subset of feature inputs as corresponding to one of the plurality of predicted settlement transaction success rates by determining a SHAP (“Shapley Additive explanation”) value for each feature.